Qwen3.8 27B repriced across 2 hosts: DeepInfra input down 50%, Ionstream cache read up 100%
Published 15 September 2026
Price move, per 1M tokens
- DeepInfra · input↓ 50%$0.40 → $0.20
- DeepInfra · output↓ 17%$3.00 → $2.50
- Ionstream · input↓ 20%$0.35 → $0.28
- Ionstream · outputunchanged$2.55 → $2.55
- Ionstream · cache read↑ 100%$0.050 → $0.100
| host | rate | was | now | change |
|---|---|---|---|---|
| DeepInfra | input | $0.40 | $0.20 | ↓ 50% |
| DeepInfra | output | $3.00 | $2.50 | ↓ 17% |
| Ionstream | input | $0.35 | $0.28 | ↓ 20% |
| Ionstream | output | $2.55 | $2.55 | — |
| Ionstream | cache read | $0.050 | $0.100 | ↑ 100% |
DeepInfra input: ↓ 50%Ionstream input: ↓ 20%Ionstream cache read: ↑ 100%
If you buy
On Ionstream, cached prompts now cost more to reuse, so re-check your cache hit rate before your next top-up if you lean on repeated context.
The model
Qwen3.8 27Bopen the model →
Open weights27.8B parameters262K context20 hosts
intelligence63rd of 168 writing93rd of 168 coding45th of 168 agents33rd of 55
Qwen3.8 27B is a downloadable model you can run yourself, and its measured strength is maths and reasoning. Agentic recovery and tool use sit near the bottom of the field, so it suits a solver rather than an agent that has to get itself back on track.
Source
Read on openrouter.ai · we published this on 15 September 2026.
read the original at openrouter.ai ↗